Главная страница Случайная страница
Разделы сайта
АвтомобилиАстрономияБиологияГеографияДом и садДругие языкиДругоеИнформатикаИсторияКультураЛитератураЛогикаМатематикаМедицинаМеталлургияМеханикаОбразованиеОхрана трудаПедагогикаПолитикаПравоПсихологияРелигияРиторикаСоциологияСпортСтроительствоТехнологияТуризмФизикаФилософияФинансыХимияЧерчениеЭкологияЭкономикаЭлектроника
The Analysis of Word Meanings
|
|
[...] All approaches to the semantic analysis of natural languages are based on the insight that the meanings of lexical items are not unanalysable or undefinable wholes. This insight has been made explicit in essentially two ways. The first is based on meaning postulates, the other on semantic components into which the lexical meanings are analysed. Meaning postulates, or semantic rules, [,,.] might be illustrated by the following examples:
1) a) boy ……… male b) girl ……… female
A rule like (1a) says that boy implies male or, what amounts to the same, that sentences like A boy is male or If x is a boy, then x is male are analytic. Meaning postulates might also involve logical contacts like ‘and’, ‘or’, ‘not’, etc.:
| 2) | a) man — male and adult |
| b) woman — female and adult | |
| c) boy or girl — not adult | |
| d) female — not male | |
| e) man or woman or boy or girl — human |
The meaning of a lexical element is therefore specified, roughly speaking, by the set of all the meaning postulates in which it occurs. More precisely: the meaning of a lexical element of L is defined implicitly by the set of all meaning postulates associated with L. [...] The second approach is that of componential analysis, which underlies the linguistic theories developed by Katz and Fodor, Weinreich, Bierwisch and others. It defines the meaning of a lexical element explicitly in terms of semantic components. These components are not parts of the vocabulary of the language itself, but rather theoretical elements, postulated in order to describe the semantic relations between the lexical elements of a given language. These components are connected again by logical constants. Thus we get the following oversimplified example:
| 3) | a) boy: animate and human and male and not adult b) girl: animate and human and female and not adult c) man: animate and human and male and adult d) woman: animate and human and female and adult |
A system of such explicitly defined lexical elements might be supplemented by a set of implicational rules of the following type:
| 4) | a) human — animate b) male — not female c) female — not male d) male — animate e) female — animate |
These implicational rules automatically complete a redundancy-free entry like (5a) to its fully specified form (5b):
| 5) | a) boy: human and male and not adult b) boy: animate and human and male and not female and not adult |
Rules of this type not only simplify the necessary dictionary specifications; they also express relevant generalizations about the semantic structure of the vocabulary described.
There is obviously a close connection between the two types of analysis just illustrated. In fact, as far as only systems of a particular kind are considered, a componential analysis of the type illustrated by (3) and (4) can be directly converted into a system of meaning postulates, and vice versa. Notice, however, that the status of elements like male in (1) and (2) is quite different from that of male in (3) and (4): while male, adult, human belong to the language described, this is not the case with respect to the components male, adult, animate, etc. The relationship between implicit and explicit definition in more complicated cases has not been systematically investigated so far. For the present exposition I will assume, quite gratuitously, that both explications are at least formally equivalent, in the sense that for each set of meaning postulates there is a componential analysis of the vocabulary involved defining the same semantic relations, and vice versa. [...]
|
|